|Safety of excluding acute pulmonary embolism based on an unlikely clinical probability by the Wells rule and normal D-dimer concentration: a meta-analysis
|Pasha SM, Klok FA, Snoep JD, Mos IC, Goekoop RJ, Rodger MA, Huisman MV
This review concluded that acute pulmonary embolism could be safely excluded in patients with an “unlikely” Wells clinical decision rule (4 or less points) and normal D-dimer test, and that anticoagulant treatment could safely be withheld. Given the potential for missed studies, clinical differences between studies and lack of power of the meta-analysis, the conclusions should be treated with caution.
To determine the negative predictive value of an “unlikely” Wells clinical decision rule (4 or less points) combined with a normal D-dimer test (detection of the cross-linked fibrin degradation fragment) to exclude pulmonary embolism, and the safety of withholding anticoagulants based on these criteria.
PubMed, EMBASE and the Cochrane Library were searched for studies published in full in English, German, French or Dutch from 2000 to December 2008; search terms were reported.
Prospective studies that evaluated a dichotomized clinical decision rule and a D-dimer test to rule out pulmonary embolism were eligible for inclusion. From these studies patients were included with an “unlikely” clinical probability and normal D-dimer test result who had not undergone any further radiological imaging aimed at evaluating the presence of pulmonary embolism or deep vein thrombosis. Studies had to have follow-up of at least three months.
The primary outcomes were confirmed fatal and non-fatal venous thromboembolism (i.e. pulmonary embolism or deep vein thrombosis). Secondary outcomes were total venous thromboembolism-attributable mortality.
In included studies, a range of D-dimer tests were used including VIDAS assay, SimpliRED, Tinaquant assay, and the Accuclot latex agglutination D-dimer test. Where reported, the cut-off for the D-dimer test was 500ng/mL. Where reported, the mean age of participants ranged from 51 to 53.5 years; the proportion of males ranged from 34.2 to 42.6%. Between 9.4 and 14.5% of participants had a history of venous thromboembolism, and between 5.7 and 18.5% had recent surgery, immobilisation or trauma (where reported). Most participants were assessed in outpatient departments.
The authors did not state how many reviewers selected studies for the review.
Assessment of study quality
Study quality was assessed in terms of patient enrolment, outcome assessment, duration of follow-up, loss to follow-up and funding source; studies had to enrol patients consecutively to be included.
The authors did not state how many reviewers performed the quality assessment.
The proportion of patients experiencing a pulmonary embolism, deep vein thrombosis or who died during follow-up, in whom acute pulmonary embolism was ruled out, was extracted. Primary authors were contacted for missing data.
Two reviewers independently extracted data; disagreements were resolved by a third reviewer.
Methods of synthesis
Weighted pooled incidences (WPIs) and negative predictive values (NPVs), with 95% confidence intervals (CIs), were calculated using both random-effects (using DerSimonian-Laird weights) and fixed-effect (using inverse arcsine variance weights) models. Heterogeneity was assessed using the I2 statistic. Patients who were lost to follow-up were excluded from the analysis. Analyses were conducted including and excluding patients receiving anticoagulants. and patients in whom additional diagnostic tests were performed despite both an “unlikely” clinical probability and a negative D-dimer. The upper limit of the 95% confidence interval of the fatal and non-fatal three months thromboembolic rate after a negative invasive pulmonary angiography was defined as the cut-off point for the safe exclusion of pulmonary embolism.
Publication bias was assessed using funnel plots.
Results of the review
Four studies met the inclusion criteria (n=1,660 patients with an “unlikely” clinical probability and normal D-dimer test, range 49 to 1,028). All studies had a follow-up of three months. Loss to follow up ranged from 0 to 0.7%. All studies were considered to be of high quality. There was no evidence of publication bias. Results presented below are for the random-effects model; the use of a fixed-effect model did significantly alter the results.
Of the 1,660 patients included in the analysis, six patients were diagnosed with symptomatic venous thromboembolism during follow-up (WPI 0.34%, 95% CI 0.036 to 0.96; I2=46.2%), and one patient died possibly as a direct consequence of a fatal pulmonary embolism (WPI 0.10%, 95% CI 0.0017 to 0.46; I2=37.2%). The pooled negative predictive value of having venous thromboembolism during three months follow-up after an “unlikely” clinical probability in combination with a normal D-dimer test was 99.7% (95% CI 99.0 to 99.9%).
When patients in whom additional diagnostic tests were performed were included (n=1,719), eight patients were diagnosed with venous thromboembolism during follow-up (WPI 0.53%, 95% CI 0.24 to 0.92), and one patient was diagnosed with fatal pulmonary embolism (WPI 0.096%, 95% CI 0.0008 to 0.42). The negative predictive value of having venous thromboembolism on follow-up was 99.5% (95% CI 99.1 to 99.8).
Acute pulmonary embolism could be safely excluded in patients with clinically suspected acute pulmonary embolism who had an “unlikely” probability and a negative D-dimer test, and anticoagulant treatment could be withheld.
The review addressed a clear research question, supported by appropriate inclusion criteria. Several relevant databases were searched, but inclusion was restricted in terms of language and publication status, so relevant studies could have been missed. Funnel plots were used to assess publication bias for some outcomes, but lack of evidence using this method did not rule out the presence of bias. Data extraction was conducted in duplicate, but it was unclear whether similar methods used to reduce error and bias were employed during study selection and the quality assessment.
Study quality was assessed, but only the broad categories assessed were given and the results for each study were not reported. There was no attempt to assess the impact of missing data on outcomes, but the rates of loss to follow-up were very low. Although there was no significant statistical heterogeneity, the studies varied in terms of the D-dimer tests used, and insufficient study and population details were provided to fully assess the degree of clinical heterogeneity across studies. Given the rarity of the outcomes being measured, the meta-analysis may have been underpowered.
Given the limitations of the review, the authors' conclusions should be treated with some caution.
Implications of the review for practice and research
Practice: The authors stated that anticoagulant therapy can safely be withheld, with no additional radiological tests, in patients with an “unlikely” Wells clinical decision rule (4 or less points) in combination with a normal D-dimer test.
Research: The authors did not state any implications for research.
Pasha SM, Klok FA, Snoep JD, Mos IC, Goekoop RJ, Rodger MA, Huisman MV. Safety of excluding acute pulmonary embolism based on an unlikely clinical probability by the Wells rule and normal D-dimer concentration: a meta-analysis. Thrombosis Research 2010; 125(4): e123-e127
Subject indexing assigned by NLM
Anticoagulants /therapeutic use; Autoimmune Diseases /complications /drug therapy /radiography; Biomarkers /blood; Cellulitis /complications /drug therapy /radiography; Coagulants; Female; Fibrin Fibrinogen Degradation Products /metabolism; Humans; Incidence; Jurisprudence; MEDLINE; Male; Middle Aged; Probability; Prospective Studies; Pulmonary Embolism /blood /diagnosis /radiography; Recurrence; Safety; Venous Thromboembolism /blood
Date bibliographic record published
Date abstract record published
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.